Recommended settings for PacBio program ccs for converting subreads to HiFi reads for downstream usage with hifiasm

[chhylp123]

Recommended settings for PacBio program ccs for converting subreads to HiFi reads for downstream usage with hifiasm

[jelber2]

Excerpt from https://genome.cshlp.org/content/suppl/2020/09/02/gr.263566.120.DC1/Supplemental_Material_.pdf (supplemental material of HiCanu paper by Nurk et al. 2020; Genome Research doi: 10.1101/gr.263566.120)

Assuming a 30-hr movie (https://www.pacb.com/products-and-services/sequel-system/latest-system-release/),
acquiring CHM13 data on four SMRTcells would require 120 hours. We obtained 3/6 NA12878 raw bam files for
the subreads and converted them to HiFi using the CCS command v3.4.1:

ccs --maxLength 21000 --minPasses 3 --numThreads 32 --polish --minPredictedAccuracy 0.99 input.bam output.bam

Would these settings be something you would recommend to start with? (Obviously adjusting threads to those available on one's specific computer).

[lh3]

I have never run PacBio's ccs tool, so I don't know.

[HenrivdGeest]

just from another user perspective; besides altering the maxLength de default settings are fine. We have good experience with the defaults, there is hardly anything thing to tweak on the ccs tools. With any alteration you can make the data just more noisy, for example lowering the minPredictedAccuracy, will just make the final result a tiny bit larger, but more noisy, which could hamper hifiasm.

[jelber2]

Oh thank you! Yeah, I guess as with most tools, there are pros/cons to changing certain settings, especially things such as k-mer length for de novo assembly!